and physics near-ir photodissociation of peroxy acetyl ... · chemistry and physics near-ir...

Atmos. Chem. Phys., 5, 385–392, 2005www.atmos-chem-phys.org/acp/5/385/SRef-ID: 1680-7324/acp/2005-5-385European Geosciences Union

AtmosphericChemistry

and Physics

Near-IR photodissociation of peroxy acetyl nitrate

S. A. Nizkorodov1, J. D. Crounse2, J. L. Fry 2, C. M. Roehl3, and P. O. Wennberg3

1Department of Chemistry, University of California at Irvine, Irvine, CA 92697, USA2Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, CA 91125, USA3Division of Geological and Planetary Sciences and Division of Engineering and Applied Science, California Institute ofTechnology, Pasadena, CA, 91125, USA

Received: 4 February 2004 – Published in Atmos. Chem. Phys. Discuss.: 1 March 2004Revised: 28 April 2004 – Accepted: 7 May 2004 – Published: 10 February 2005

Abstract. Measurements of the C-H overtone transitionstrengths combined with estimates of the photodissociationcross sections for these transitions suggest that near-IR pho-todissociation of peroxy acetyl nitrate (PAN) is less signifi-cant (Jnear−IR≈3×10−8 s−1 at noon) in the lower atmospherethan competing sinks resulting from unimolecular decompo-sition and ultraviolet photolysis. This is in contrast to thephotochemical behavior of a related peroxy nitrate, pernitricacid (PNA), that undergoes rapid near-IR photolysis in theatmosphere withJnear−IR≈10−5 s−1 at noon (Roehl et al.,2002). This difference is attributed to the larger binding en-ergy and larger number of vibrational degrees of freedom inPAN, which make 4νCH the lowest overtone excitation witha high photodissociation yield (as opposed to 2νOH in PNA).

1 Introduction

Atmospheric photochemistry is normally driven by elec-tronic state excitations in the ultraviolet range of the so-lar spectrum. However, photodissociation processes re-sulting from near-IR direct overtone excitations within theground electronic state can also be significant in special cases(Donaldson et al., 2003). Donaldson et al. (1997) exam-ined several atmospheric molecules featuring small bind-ing energies and strong overtone transitions, and identifiedHO2NO2 (pernitric acid or PNA) as the most likely candi-date for an efficient overtone photodissociation. They arguedthat the dissociation energy threshold for PNA, D0(HO2-NO2)=7970±280 cm−1 (Zabel, 1995), is sufficiently low topermit its overtone photolysis via the second OH stretchingovertone transition at E(3νOH)=10 090 cm−1. Subsequentexperimental study of near-IR photodissociation cross sec-tions of PNA by Roehl et al. (2002) showed that PNA over-

Correspondence to:S. A. Nizkorodov([email protected])

tone dissociation is even more important than Donaldson etal. (1997) initially surmised. Roehl et al. (2002) found thatPNA can photolyze with a substantial yield not only upon3νOH but also upon 2νOH excitation at 6900 cm−1, whichlies as much as 1070 cm−1 below the PNA formal dissoci-ation energy threshold. These striking results have recentlybeen independently confirmed by Sinha’s research group(Matthews et al., 2004). Such efficient under-threshold pho-tolysis can be attributed to the collisional activation and inter-nal energy contributions to the dissociation process, just as inthe well-known cases of ultraviolet photodissociation of NO2(Pitts et al., 1964; Roehl et al., 1994) and O3 (Finlayson-Pitts and Pitts, 2000), and overtone dissociation of nitric acid(Sinha et al., 1990). The combined 2νOH and 3νOH PNAatmospheric photodissociation rate amounts to≈10−5 s−1,making near-IR photolysis the dominant sink process forPNA in the upper troposphere and lower stratosphere (Roehlet al., 2002).

This letter examines possible atmospheric significanceof near-IR photolysis for another important peroxy ni-trate, CH3C(O)OONO2 (peroxy acetyl nitrate or PAN). PANis a key odd-nitrogen component of photochemical smog(Finlayson-Pitts and Pitts, 2000), and it serves both as a tem-porary reservoir and as a global transporter for NOx. PANoften accounts for the largest fraction of all odd-nitrogencompounds (NOy) in the atmosphere, especially at high lat-itudes (Ridley et al., 2003; Stroud et al., 2003). The pri-mary known sinks for PAN include unimolecular decomposi-tion, which is the dominant channel in the lower atmosphere,and ultraviolet photolysis, which becomes prevalent in thecolder upper tropospheric layers. Reaction of PAN with OHis too slow to be of relevance (DeMore et al., 1997). Sim-ilar to other peroxy nitrates (Kirchner et al., 1999; Zabel,1995), PAN has a low dissociation energy, D0(CH3C(O)OO–NO2)=9820±240 cm−1 (Bridier et al., 1991; Zabel, 1995),which falls between the 3νCH (E≈8550 cm−1) and 4νCH(E≈11 100 cm−1) vibrational states of the molecule. To

© 2005 Author(s). This work is licensed under a Creative Commons License.

386 S. A. Nizkorodov et al.: Near-IR photodissociation of peroxy acetyl nitrate

evaluate the efficiency of near-IR photochemistry of PAN,knowledge of both the intensity of the CH stretching over-tones and photodisssociation quantum yields are required.Although the strength of CH stretching overtone transitionsare well-known for many hydrocarbons (Hanazaki et al.,1985; Kjaergaard et al., 1991; Nakagaki and Hanazaki,1986), no information on the cross sections and photodis-sociation quantum yields for PAN is presently available. Inwhat follows, we show that these quantities can be estimatedfrom statistical models and comparisons with experimentaldata on related molecules, and we use them to predict thenear-IR photodissociation rates for PAN under typical atmo-spheric conditions.

2 Results and discussion

2.1 Definitons and units

Subject to the approximations described below, band-specificatmospheric photodissociation rates,Jband, can be calculatedin the following way:

Jband=

∫band

F(v̄)σ (v̄)φ(v̄)dv̄≈ 〈F(v̄)〉 ×Sband×φband (1)

F(v̄) is the frequency-dependent solar flux measured inphotons

cm2s cm−1 , 〈F(v̄)〉 is the average value ofF(v̄) over the ab-sorption band,σ(v̄) is base-e molecular absorption crosssection in cm2/molecule,φ(v̄) is photodissociation quantumyield, v̄ is wave number in cm−1, andSband=

∫band

σ(v̄)dv̄ is

the integrated absorption cross section (in cm/molecule) forthe band in question. The last equality applies only ifF(v̄)

varies weakly with frequency, which is usually a good ap-proximation for the narrow near-IR bands unless they areclose to strong water or carbon dioxide absorptions. Thequantityφband is best interpreted as the ratio between the in-tegrated photodissociation and absorption cross sections forthe band. The paragraphs below deal with estimation ofS

andφ for the CH stretching overtone transitions of PAN.

2.2 PAN overtone absorption cross-sections

Integrated absorption cross sections for PAN fundamentalshave been reported by Gaffney et al. (1984). However, theiranalysis has subsequently been questioned and corrected byTsalkani and Toupance (1989) for the four strongest infraredbands of PAN below 2000 cm−1. To resolve the resultingambiguity of the intensity data for the remaining bands notconsidered by Tsalkani and Toupance (1989), we have mea-sured relative intensities of the CH stretching fundamentaland overtone transitions and determined absolute transitionintensities by relation to Tsalkani and Toupance (1989) datafor NO stretching vibrations of PAN. Pure gaseous PAN sam-ples were synthesized as described by Gaffney et al. (1984),

and gas-phase spectra of PAN were recorded with an FTIRspectrometer at 1 cm−1 resolution (Fig. 1). Depending on theexamined spectral range, we used either a single-pass 20 cmcell containing 0.1–1 Torr partial pressure of PAN, or a mul-tipass cell with an effective path length of 360 cm containingthe equilibrium vapor pressure of PAN at room temperature(about 4.5 Torr). The multipass configuration was requiredto observe the second CH stretching overtone with an ac-ceptable signal-to-noise ratio (Fig. 1).

The ratios of PAN band strengths were ob-tained from multiple spectra. The absorbances(A=σ(v̄)×[PAN]×pathlength) for the bands in ques-tion were always kept small to avoid deviations from theBeer-Lambert law. For very weak transitions like CH-overtones there is no reason to believe that their integratedabsorbances should deviate from a linear dependence onPAN concentration. For the stronger fundamental transi-tions, deviations from linearity are certainly possible, but wemeasured the ratio of the integrated absorbances at multiplepartial pressures of PAN and always obtained the same result.This is an indirect proof that Beer-Lambert law was obeyedfor these transitions under our experimental conditions. Thevalue ofS(νCH) was measured relative to the band strengthsof NO2 stretches reported by Tsalkani and Toupance (1989).Our value,S(νCH)=1.57×10−18 cm/molecule, is 40% lowerthan that reported by Gaffney et al. (1984). This agreeswith the conclusions of Tsalkani and Toupance (1989), whoshowed that data from Gaffney et al. (1984) overestimatedband intensities of PAN by 5 to 35%. The CH stretchingovertone band strengths were always measured relativeto the closest available CH stretching transition at lowerfrequency. Our resultsS(2νCH)/S(νCH)=0.096±0.015 andS(3νCH)/S(2νCH)=0.085±0.030 are consistent with thegenerally observed trend of the 10-fold drop of overtoneintensity per vibrational quantum (we defineS(2νCH) andS(3νCH) as the total integrated absorption cross sectionfor the 5600–6300 cm−1 and 8300–9000 cm−1 ranges,respectively).

Integrated intensity for the 4νCH overtone of PAN couldnot be reliably determined from our FTIR spectra, but it canbe deduced from comparison with spectra of other moleculesbearing an acetyl group. Indeed, the observed spectra of thefirst and second CH overtones of PAN (Fig. 1) closely re-semble the spectra of acetone and acetaldehyde (Hanazakiet al., 1985). The orientational site splitting pattern arisingfrom the inequivalence of CH bonds in the methyl group(Hanazaki et al., 1985; Kjaergaard et al., 1991) is similar forall three molecules. The integrated intensities (per methylgroup) for theνCH fundamental, 2νCH first overtone, and3νCH second overtone bands are also quite similar (Table 1).Assuming that the PAN overtone intensity decreases with vi-brational quantum number by the same factors as the aver-age of the corresponding acetone and acetaldehyde values,S(4νCH)/ S(3νCH)=0.059, we can estimateSPAN (4νCH) as7.6×10−22 cm/molecule. Based on the good uniformity of

Atmos. Chem. Phys., 5, 385–392, 2005 www.atmos-chem-phys.org/acp/5/385/

S. A. Nizkorodov et al.: Near-IR photodissociation of peroxy acetyl nitrate 387

Fig. 1. IR spectrum of PAN recorded in this work:(a) fundamental range displaying well-known PAN bands;(b) CH stretching fundamentalrange;(c) first CH stretching overtone range; d) second CH stretching overtone range. The CH-stretching overtones have a complicated bandstructure resulting from the inequivalence of CH bonds in the methyl group. Experimentally measured ratios between integrated absorptioncross sections areS(νCH):S(2νCH):S(3νCH) = 1:0.096:0.0082.S(νCH) is measured to be 1.57×10−18 cm/molecule (basee) in this work.

the experimentally available band strength ratios for thesethree molecules, we do not expect the estimated value ofSPAN (4νCH) to deviate from the true band strength by morethan a factor of two.

2.3 Photodissociation quantum yields

Photodissociation quantum yields for CH overtone excita-tions of PAN have been calculated with the Multiwell suite(Barker, 2001) using a statistical approach. The calculationinvolves shifting the Boltzmann internal energy distributionof the molecule upward by the photon energy, and solvingthe coupled master equations for discretized internal energystates. Excited molecules can undergo either a unimoleculardecomposition with the microcanonical rate constant kuni(E)or collision-induced internal energy change with the rate thatdepends on pressure and internal energy. The fraction ofmolecules that end up decomposing by the time internal en-ergy in the surviving molecules is nearly thermalized (≈103

collisions) is taken as the photodissociation yield.The quantum yield calculation presented here is based on

the assumptions that: i) the internal energy can exchangefreely between different vibrational degrees of freedom (K-rotor states are also counted towards the total available en-ergy); ii). The transition strengths for sequence (hot) bandsoriginating from the low- frequency vibrational states andbuilding upon CH overtones are the same as for the cor-responding overtones themselves. The high frequency CHvibrations are known to couple quite weakly with the low-frequency vibrational modes. Therefore, the second assump-tion is likely to hold, whereas the first one may in princi-

Table 1. Experimental integrated absorption cross sections (basee) for the CH-stretching overtones of acetone, acetaldehyde, andPAN. The units ofS(band) are cm2/molecule cm−1 (1 cm−2 atm−1

at 298 K=4.063×10−20cm/molecule). The integrated absorptioncross sections are given per CH3 group; i.e. the acetone values havebeen divided by 2.

Band Acetone/2a Acetaldehydea PAN

(νCH) 3.04E-18 2.02E-18 1.57E-182(νCH) 1.28E-19 1.51E-19 1.51E-19b

3(νCH) 1.60E-20 1.63E-20 1.28E-20c

4(νCH) 1.09E-21 8.21E-22 7.56E-22d

a From Hanazaki et al. (1985).b Measured to be 9.6% ofS(νCH) in this work.c Measured to be 8.5% ofS(2νCH) in this work.d Assumed value based on comparison with acetone and acetalde-hyde (5.9% of S(3νCH)).

ple break down leading to non-statistically long lifetimes ofthe vibrationally excited molecules. At finite pressures, thedissociation quantum yield may be reduced by these non-statistical effects. The yields we are predicting will there-fore represent an upper limit to the actual photodissociationyields.

The exponential collision model built into the Multi-well was used, with the probability of downward collisionstaken ase(Efinal−Einitial)/1, where1=45 cm−1+0.005×Einitial ,a weak collision model (Barker, 2001). Vibrational

www.atmos-chem-phys.org/acp/5/385/ Atmos. Chem. Phys., 5, 385–392, 2005


Fig. 2. Experimental photodissociation quantum yields for differ-ent vibrational overtone excitations of PNA measured by Roehlet al. (2002) (filled circles) compared to the statistical simulationresults (open squares) of this work. Both simulations and datacorrespond to P=5 Torr. The open circle point in the 2ν1 panelcomes from Matthews et al. (2004) and it corresponds to T=298 K,P=0.1 Torr. Parameters for our statistical simulation were chosen toreproduce the 2ν1 photodissociation yields.

frequencies for PAN were taken from Miller et al. (1999).Lennard-Jones collision parameters were the same as usedby Bridier et al. (1991). The NO2 and –C-O-O moieties weretreated as hindered rotors in the transition state. The CH3group was treated as a free rotor for both reactant and tran-sition state. The initially-selected RRKM parameters werefine-tuned to reproduce i) high pressure limit of the experi-

mental unimolecular decomposition rate (Bridier et al., 1991;Zabel, 1995); ii) the equilibrium constant for PAN decom-position (Bridier et al., 1991) at all experimentally availabletemperatures. The final choice of parameters for the calcula-tion could reproduce both experimental data sets within 30%.

The predictive power of the calculation was first testedon PNA because photodissociation quantum yields have al-ready been measured for this molecule (Roehl et al., 2002).Figure 2 compares the experimental results with our sta-tistical calculations for theνOH+2νOOH combination band,2νOH overtone, and 2νOH+νOOH combination band of PNA(listed here in the order of increasing excitation energy).Required RRKM parameters for PNA were taken from ex-periment and from Chen and Hamilton (1996) and Zabel(1995). The simulation reproduces both the magnitude andthe temperature dependence of the photodissociation quan-tum yields with good accuracy. To achieve this level ofagreement, we had to reduce the dissociation energy thresh-old of PNA by 200 cm−1 from the accepted experimen-tal value, which is well within the error bars for D0(HO2–NO2)=7970±280 cm−1 (Zabel, 1995). Staikova et al. (2002)also reported a lower dissociation energy for PNA.

A curious side result of this simulation is the prediction ofunity quantum yield for the 2νOH+νOOH excitation of PNA.The experimental data suggest a yield that is slightly belowunity for this transition. This discrepancy may be a reflec-tion of not entirely statistical behavior of PNA at these ex-citation energies. It may also reflect deficiencies of the cho-sen weak collision model. However, the experimental datafor 2νOH+νOOH andνOH+2νOOH bands are relatively noisy(Fig. 2) making it difficult to draw more definite conclusions.Matthews et al. (2004) assumed a photodissociation quantumyield of one for the 2νOH+νOOH band in their work.

The predicted dissociation yields for PAN excitation at8700 cm−1 (i.e. within the blue wing of the 3νCH band) areshown in Fig. 3 for several atmospherically relevant temper-atures. The most interesting observation is that the photodis-sociation yields for PAN display a remarkably broad pres-sure dependence. In contrast, the photodissociation yield forthe 2νOH overtone transition of PNA is fairly constant be-low P=1 atm. This important difference can be attributedto the substantially larger number of vibrational degrees offreedom and larger binding energy in PAN resulting in lowermicrocanonical decomposition rates. The collisional deac-tivation and excitation rates can, therefore, easily competewith unimolecular decomposition even when the system isplaced above the dissociation energy. In fact, the calcu-lated PAN dissociation yields do not reach unity under typi-cal atmospheric pressures even for excitation lying as muchas 1000 cm−1 above the dissociation threshold (Fig. 4). Incontrast, PNA is predicted to fall apart with 100% efficiencywhen excited so high above the dissociation energy thresholdat pressures below one atmosphere.



Table 2. Calculated photodissociation quantum yields for PAN overtone transitions under representative atmospheric conditions. TheJnear−IR column lists the total near-IR photodissociation rates (J3νCH+J4νCH) calculated from the integrated absorption cross sections fromTable 1 and a solar flux of 5×1014photons cm−2 s−1 nm−1. The last column is the fractional contribution of under-threshold dissociationvia 3νCH to the total rate (i.e.J3νCH/Jnear−IR).

Altitude T P φ(3νCH) φ(4νCH) Jnear−IR 3νCH contribution(km) (K) (Torr) (s−1) (%)

0 288 760 0.0041 0.80 2.8E-8 14.83 269 526 0.0027 0.82 2.8E-8 9.36 249 354 0.0016 0.85 2.7E-8 5.59 230 230 0.0009 0.88 2.8E-8 3.012 217 143 0.0007 0.92 2.9E-8 2.215 217 90 0.0011 0.97 3.1E-8 3.1

Fig. 3. Pressure dependence of the predicted photodissociationquantum yields for 8700 cm−1 excitation of PAN (representative ofthe 3νCH excitation) at different temperatures. Calculated photodis-sociation yields for 2νOH band of PNA at 273 K are also shown forcomparison. In contrast to the PNA case, photodissociation quan-tum yields for PAN photolysis at 8700 cm−1 strongly depend onpressure under typical atmospheric conditions.

2.4 near-IR photodissociation rates for atmospheric PAN

Simulated PAN photodissociation yields for a standardizedatmosphere are given in Table 2. The effective photodis-sociation yield for 3νCH band (estimated as an average ofyields calculated for 8450, 8550, and 8650 cm−1 excitations)is very small compared to the 4νCH yield (corresponding to11 100 cm−1 excitation). Using the transition strength datafrom Table 1 and a typical near-IR solar flux of 5×1014 pho-tons cm−2 s−1 nm−1 (Finlayson-Pitts and Pitts, 2000), weestimate the total near-IR photodissociation rate for PANasJnear−IR≈3×10−8 s−1 with a fairly weak altitude depen-dence (Table 2).

Because of the very low contribution of the 3νCH band ofPAN to the total near-IR photolysis rate (<5% above 6 km),

Fig. 4. Pressure dependence of the predicted photodissociationquantum yields of PAN at 298 K at different excitation frequen-cies. The frequencies cover the range 8400 cm−1 (bottom curve)to 11 700 cm−1 (top curve) in 300 cm−1 increments. Even at thehighest excitation energy, the yield is still noticeably different fromunity at typical atmospheric pressures.

the predicted rate is relatively insensitive to the calculatedphotodissociation yields. For example, an order of magni-tude increase inφ(3νCH) would translate to less than 50%increase in the total photolysis rate. Theφ(4νCH) values arealready above 0.8; adjusting them to the maximal value ofunity would only produce a marginal increase in the rate. Theconfidence in the predicted rate is, therefore, dominated bythe uncertainty inS(4νCH) believed to be accurate to withina factor of two (Table 1). Given these uncertainties, a veryconservative upper limit for the near-IR PAN photodissocia-tion rate under typical atmospheric conditions can be set atJnear−IR(PAN) <10−7 s−1. Using similar reasoning, one canalso show thatJnear−IR is insensitive to poor atmospherictransmittance at wavelengths corresponding to 3νCH transi-tion (the 4νCH spectral range is clean).



Fig. 5. Dependence of the predicted 24 hour-averaged near IR pho-todissociation rates (thick solid lines), UV photolysis rates (dashedlines), and unimolecular decomposition rate (thin solid line) on alti-tude. The UV photolysis rates were calculated for actual conditionsof 8 May, 67◦ N (short dash) and 25 September, 35◦ N (long dash).Two near IR photolysis curves correspond to 18-hour (8 May) and12-hour (25 September) day lengths with an average solar flux of5×1014photons cm−2 s−1 nm−1. The near IR channel is seen toplay a minor role (<10%) at all altitudes.

This limit is to be contrasted with a characteristic near-IR photolysis rate for PNA,Jnear−IR(PNA)≈10−5 s−1, whichis higher by two orders of magnitude. The large differ-ence in the photodissociation rates for PAN and PNA re-flects the disparity in absorption cross sections for the over-tone excitations responsible for their near-IR photodissoci-ation activity. Whereas the under-threshold 2νOH excita-tion accounts for the major fraction ofJnear−IR(PNA), the3νCH band of PAN is but a minor contributor (<5% above6 km) to Jnear−IR(PAN). Because the overtone absorptionstrengths go down by roughly one order of magnitude witheach vibrational quantum, the jump in two vibrational quantabetween the primary photolytic excitations in PAN(4νCH)

and PNA(2νOH) can be held responsible for the observeddifference in theJnear−IR values. In addition, CH3 over-tone intensities in acetyl groups are known to be systemat-ically lower than those of other kinds of CH3 groups (Nak-agaki and Hanazaki, 1986). They are also considerablylower than intensities of typical OH stretching overtones (e.g.compareSPNA(2νOH)=(9.5±1.9)×10−19 cm/# (Roehl et al.,2002) withSPAN(2νCH)=1.51×10−19 cm/#).

Although the predicted near-IR photodissociation rate ofPAN is quite low (the 24-h average of 1.5×10−8 s−1 trans-lates into an atmospheric lifetime of 770 days), it becomesstrongly competitive with the thermal decomposition rate

(Zabel, 1995) above 6 km (Table 2 and Fig. 5). It is alsohigher than the upper limit for the rate of OH+PAN re-action (kOH+PAN[OH]<10−8 s−1 for the troposphere; De-More et al., 1997). Comparison with the UV photolysis rateof PAN is less straightforward because UV radiation is at-tenuated by the atmosphere to a much greater extent thannear-IR at high solar zenith angle (SZA). For example, at15 km altitude ,JUV (PAN) ≈6×10−7 s−1 at SZA=0◦ butit drops to≈6×10−8 s−1 at SZA=86◦ (solar fluxes for thiscalculation are taken from Finlayson-Pitts and Pitts, 2000and PAN absorption cross sections are from DeMore et al.,1997). On the contrary, the estimated near-IR photolysisrate (≈3×10−8 s−1) is reduced by only 35% upon this SZAchange. Consequently, the near-IR photolysis of PAN canpotentially become comparable to the UV photolysis rate atvery high SZA, providing a small contribution to the twilightproduction of HOx radicals at high altitudes (Wennberg et al.,1999).

Figure 5 displays 24-hour averaged rates of UV photoly-sis, near-IR photolysis, and thermal decomposition for PAN.The UV photolysis rates correspond to conditions foundat high and middle latitudes near equinox as described inSalawitch et al. (2002) (8 May 1997 with latitudes between65◦ N and 70◦ N and 25 September 1993 at 35◦ N). Near-IR rates are estimated by assuming day lengths of 18 h and12 h, respectively, and a near-IR solar flux of 5×1014 pho-tons cm−2 s−1 nm−1. The thermal decomposition rate forPAN is calculated from data of Zabel (1995). Figure 5shows that< Jnear−IR(PAN)> is approximately 5–10% of<JUV(PAN)> at high tropospheric altitudes. At lower al-titudes, both photolysis channels are dwarfed by the ther-mal decomposition. We conclude that near-IR photolysis isnot likely to significantly affect the average atmospheric life-time of PAN. Given that it is hard to quantify the uncertain-ties of the predicted photodissociation yields and absorptioncross sections of PAN, however, the near-IR photolysis ofPAN in the atmosphere cannot be completely discounted atthis stage. More experimental research in this area is clearlyneeded.

2.5 Experiments on near-IR photolysis of PAN

We have attempted to measure photodissociation cross sec-tions of PAN in the 3νCH–4νCH range using the action spec-troscopy approach similar to that used for PNA by Roehl etal. (2002). PAN/N2 mixture at 1–5 Torr and 250–300 K wasirradiated with a tunable pulsed near-IR laser in the presenceof a small amount of NO and O2. A dye laser was employedin a fluorescence scheme to monitor the OH produced in themixture following the photolysis laser pulse. If the near-IRlaser were to photolyze PAN molecules in the gas flow, amomentary increase in the OH fluorescence signal would beobserved assuming that the following reaction sequence runsto completion:

PAN + hν(near−IR) → CH3C(O)O2 + NO2



CH3C(O)O2 + NO → CH3 + CO2 + NO2

CH3 + O2 → CH3O2

CH3O2 + NO → CH3O + NO2

CH3O + O2 → CH2O + HO2

HO2 + NO → OH + NO2.

Despite numerous attempts under various reaction condi-tions, we have not been able to observe any near-IR laserinduced increase in the OH signal. Detailed kinetic simula-tions of our experimental conditions predict a signal-to-noiseratio reduction of≈102 for 3νCH band of PAN compared to3νOH band of PNA (which could be successfully observed inthis apparatus, Roehl et al., 2002) assuming unity photodis-sociation yield in both cases. This large penalty is comprisedof i) a factor of 2 reduction in the integrated absorption crosssection; ii) factor of 5 increase in the effective band width forthe CH3 vs. OH second overtone; iii) a factor of∼10 reduc-tion in the efficiency of conversion of CH3C(O)O2 into OHin the chemical sequence above, as opposed to that of HO2into OH in the case of PNA photodissociation. The signal-to-noise ratio in our best 3νOH photodissociation spectrum ofPNA (≈103, Roehl et al., 2002) allows us to place an upperlimit of 10% on the photodissociation yield of the 3νCH stateof PAN at 1–5 Torr and 250–300 K, which is in rough agree-ment with simulations presented here. The predicted signal-to-noise ratio for the 4νCH band of PAN is<0.5, consistentwith the failure to experimentally detect it.

We are currently designing an experiment that will allowus to measure CH3C(O)O2 following pulsed near IR laser ex-citation of PAN/N2 mixtures using a CIMS (chemical ioniza-tion mass spectrometry) technique. Preliminary tests showpositive results using UV photons to dissociate PAN. The re-sults of this work will be published elsewhere.

3 Conclusions

Due to its unique combination of binding energy and molec-ular size, PNA may well turn out to be the only peroxy nitratethat is controlled by near-IR photolysis (Roehl et al., 2002) inthe atmosphere. The larger binding energies and higher de-grees of molecular complexity make peroxy acetyl nitrates,like PAN, considerably more resistant to direct overtone pho-tolysis in the near-IR range of the solar spectrum. Specifi-cally, our analysis predicts a near-infrared photodissociationrate of≈3×10−8 s−1 for PAN at noon, largely independentof altitude. The near-IR photolysis of PAN can become com-petitive with UV photolysis under twilight illumination con-ditions, but even in the coldest upper tropospheric layers it iscalculated to contribute only 5–10% to the 24-hour averagedremoval rate of PAN.

Acknowledgements.This work was funded in part by NASA’sAtmospheric Effects of Aviation Program and the AtmosphericChemistry program of the National Science Foundation. SANthanks the Camille and Henry Dreyfus Foundation for financialsupport during his postdoctoral research at Caltech. The authorsthank R. Salawitch (JPL) for providing 24-averaged UV-photolysisrates of PAN used in Fig. 5 and A. Sinha (UCSD) for a preprint ofRef. (Matthews et al., 2004).

Edited by: J. Abbatt

References

Barker, J. R.: Multiple-well, multiple-path unimolecular reactionsystems. I. Multiwell computer program suite, Int. J. Chem. Ki-netics, 33, 232–245, 2001.

Bridier, I., Caralp, F., Loirat, H., Lesclaux, R., Veyret, B., Becker,K.H., Reimer, A. and Zabel, F.: Kinetic and theoretical studiesof the reactions CH3C(O)O2+NO2+M ↔ CH3C(O)O2NO2+Mbetween 248 K and 393 K and between 30 Torr and 760 Torr, J.Phys. Chem., 95, 9, 3594–3600, 1991.

Chen, Z. and Hamilton, T. P.: Ab initio study of pernitric acid:comparison with experimental spectra, J. Phys. Chem., 100, 39,15 731–15 734, 1996.

DeMore, W. B., Sander, S. P., Golden, D. M., Hampson, R. F.,Kurylo, M. J., Howard, C. J., Ravishankara, A. R., Kolb, C. E.,and Molina, M. J.: Chemical kinetics and photochemical data foruse in stratospheric modeling: Evaluation number 12. Jet Propul-sion Laboratory, Pasadena, 1997.

Donaldson, D. J., Frost, G. J., Rosenlof, K. H., Tuck, A. F., andVaida, V.: Atmospheric radical production by excitation of vi-brational overtones via absorption of visible light, Geophys. Res.Lett., 24, 21, 2651–2654, 1997.

Donaldson, D.J., Tuck, A.F. and Vaida, V.: Atmospheric Photo-chemistry via vibrational overtone absorption, Chem. Rev., 103,12, 4717–4729, 2003.

Finlayson-Pitts, B. J. and Pitts, J. N.: Chemistry of the upper andlower atmosphere: theory, experiments, and applications, Aca-demic Press, 1040, 2000.

Gaffney, J. S., Fajer, R., and Senum, G. I.: An improved proce-dure for high purity gaseous peroxyacyl nitrate production: useof heavy lipid solvents, Atmos. Environ., 18, 1, 215–18, 1984.

Hanazaki, I., Baba, M., and Nagashima, U.: Orientational site split-ting of methyl C-H overtones in acetone and acetaldehyde, J.Phys. Chem., 89, 26, 5637–5645, 1985.

Kirchner, F., Mayer-Figge, A., Zabel, F., and Becker, K. H.: Ther-mal stability of peroxynitrates, Int. J. Chem. Kinetics, 31, 2, 127–144, 1999.

Kjaergaard, H. G., Henry, B. R., and Tarr, A. W.: Intensities in localmode overtone spectra of dimethyl ether and acetone, J. Chem.Phys., 94, 9, 5844–5854, 1991.

Matthews, J., Sharma, R., and Sinha, A.: Photodissociation of vi-brationally excited pernitric acid: HO2NO2 (2vOH) + 390 nm, J.Phys. Chem. A, 108, 39, 8134–8139, 2004.

Miller, C. E., Lynton, J. I., Keevil, D. M., and Francisco, J. S.:Dissociation pathways of peroxyacetyl nitrate (PAN), J. Phys.Chem., 103, 51, 11 451–11 459, 1999.

Nakagaki, R. and Hanazaki, I.: Overtone spectroscopy of acetylcompounds. Inequivalent carbon-hydrogen (CH) oscillators and



substituent effects of the carbonyl group, Chem. Phys. Lett., 128,4, 432–438, 1986.

Pitts Jr., J. N., Sharp, J. H., and Chan, S. I.: Effects of wavelengthand temperature on primary processes in the photolysis of ni-trogen dioxide and a spectroscopic-photochemical determinationof the dissociation energy, J. Chem. Phys., 40, 12, 3655–3662,1964.

Ridley, B. A., Atlas, E. L., Montzka, D. D., Browell, E. V., Cantrell,C. A., Blake, D. R., Blake, N. J., Cinquini, L., Coffey, M. T., Em-mons, L. K., Cohen, R. C., DeYoung, R. J., Dibb, J. E., Eisele, F.L., Flocke, F. M., Fried, A., Grahek, F. E., Grant, W. B., Hair, J.W., Hannigan, J. W., Heikes, B. J., Lefer, B. L., Mauldin, R. L.,Moody, J. L., Shetter, R. E., Snow, J. A., Talbot, R. W., Thornton,J. A., Walega, J. G., Weinheimer, A. J., Wert, B. P., and Wim-mers, A. J.: Ozone depletion events observed in the high-latitudesurface layer during the TOPSE aircraft program, J. Geophys.Res. D, 108, D4, TOP 4/1-TOP 4/22, 2003.

Roehl, C. M., Nizkorodov, S. A., Zhang, H., Blake, G. A., andWennberg, P. O.: Photodissociation of peroxynitric acid in thenear-IR, J. Phys. Chem. A, 106, 15, 3766–3772, 2002.

Roehl, C. M., Orlando, J. J., Tyndall, G. S., Shetter, R. E., Vazquez,G. J., Cantrell, C. A., and Calvert, J. G.: Temperature depen-dence of the quantum yields for the photolysis of NO2 near thedissociation limit, J. Phys. Chem., 98, 32, 7837–7843, 1994.

Salawitch, R. J., Wennberg, P. O., Toon, G. C., Sen, B., and Blavier,J.-F.: Near IR photolysis of HO2NO2: implications for HOx,Geophys. Res. Lett., 29, 16, 9/1–9/4, 2002.

Sinha, A., Vander Wal, R.L. and Crim, F.F.: State-resolved uni-molecular reactions: the vibrational overtone initiated decompo-sition of nitric acid, J. Chem. Phys., 92, 1, 401–10, 1990.

Staikova, M., Donaldson, D.J. and Francisco, J.S.: Overtone-induced reactions on the HO2NO2 potential surface, J. Phys.Chem. A, 106, 12, 3023–3028, 2002.

Stroud, C., Madronich, S., Atlas, E., Ridley, B., Flocke, F., Wein-heimer, A., Talbot, B., Fried, A., Wert, B., Shetter, R., Lefer, B.,Coffey, M., Heikes, B., and Blake, D.: Photochemistry in thearctic free troposphere: NOx budget and the role of odd nitrogenreservoir recycling, Atmos. Environ., 37, 24, 3351–3364, 2003.

Tsalkani, N. and Toupance, G.: Infrared absorptivities and inte-grated band intensities for gaseous peroxyacetyl nitrate (PAN),Atmos. Environ., 23, 8, 1849–1854, 1989.

Wennberg, P. O., Salawitch, R. J., Donaldson, D. J., Hanisco, T. F.,Lanzendorf, E. J., Perkins, K. K., Lloyd, S. A., Vaida, V., Gao,R. S., Hintsa, E. J., Cohen, R. C., Swartz, W. H., Kusterer, T.L., and Anderson, D. E.: Twilight observations suggest unknownsources of HOx, Geophys. Res. Lett., 26, 10, 1373–1376, 1999.

Zabel, F.: Unimolecular Decomposition of Peroxynitrates, Z. Phys.Chem., 188, 119–142, 1995.


and physics near-ir photodissociation of peroxy acetyl ... · chemistry and physics near-ir...

Documents